To determine the optimal dose of radiosurgery-alone for patients with 1-3 cerebral metastases from breast cancer.Patients receiving 20 Gy (n=20) were compared to those receiving 16-18.5 Gy (n=10) for local control, distant brain control and overall survival. Seven other variables were also evaluated.Radiosurgery dose achieved significance on univariate (p=0.002; log-rank and Wilcoxon test) and multivariate analysis (p=0.004) of local control. Twelve-month local control rates were 94% after 20 Gy and 48% after 16-18.5 Gy. On univariate analysis of distant brain control, radiosurgery dose was not a significant factor, with 12-month rates of 73% and 60 respectively. Regarding overall survival, radiosurgery dose was of borderline significance (p=0.059; Wilcoxon test). Twelve-month overall survival rates were 75% and 40 respectively. On Cox regression analysis, radiosurgery dose exhibited a trend for improving survival (p=0.10).Radiosurgery with 20 Gy resulted in significantly better local control and led to a trend towards improved overall survival compared to treatment with 16-18.5 Gy.